@agentic-name-service/sdk
FrameworkFreeOfficial Agent SDK for the Agentic Name Service (ANS) — orchestrates MCP tool calls across Gateway and Guardian for trilateral authentication
Capabilities13 decomposed
trilateral-agent-authentication-orchestration
Medium confidenceOrchestrates a three-party authentication handshake between an Agent, Gateway, and Guardian using MCP (Model Context Protocol) tool calls. The SDK manages the state machine for credential exchange, signature verification, and mutual authentication across distributed parties without centralizing trust. Implements a request-response pattern where the Agent initiates calls through the Gateway, which delegates verification to the Guardian, then returns authenticated tokens back through the MCP channel.
Implements a trilateral handshake pattern specifically designed for MCP tool calls, where authentication state is managed across three independent parties without a central authority. Uses MCP's native tool registry to define authentication endpoints, avoiding custom protocol definitions.
Differs from OAuth2/OIDC by eliminating the central authorization server and distributing trust across Gateway and Guardian; differs from mutual TLS by operating at the application layer within MCP, allowing agent-level granularity.
mcp-tool-call-routing-with-auth-context
Medium confidenceRoutes MCP tool calls from agents through the Gateway with automatic injection of authentication context derived from the trilateral handshake. The SDK wraps the native MCP tool invocation mechanism, intercepts calls, appends signed auth tokens to the request metadata, and validates response signatures before returning results to the agent. Implements a middleware pattern that is transparent to the agent's tool-calling logic.
Implements authentication as a transparent middleware layer within the MCP tool-calling pipeline, using MCP's native metadata mechanism rather than custom headers. Signature verification happens on response, not just request, ensuring bidirectional trust.
More lightweight than API gateway solutions like Kong because it operates at the SDK level without requiring a separate infrastructure component; more flexible than hardcoded auth headers because it derives credentials from the active session state.
audit-logging-of-authentication-events
Medium confidenceRecords detailed audit logs of all authentication events (credential submission, Guardian verification, token generation, session renewal, expiration) with timestamps, actor information, and outcomes. The SDK provides structured logging hooks that emit events at each stage of the trilateral handshake, allowing downstream systems to capture and analyze authentication activity. Implements a pluggable logger interface that supports multiple backends (console, file, remote syslog, cloud logging services).
Provides pluggable audit logging at each stage of the trilateral handshake with structured event format, allowing organizations to integrate authentication events into their existing logging and monitoring infrastructure. Includes built-in redaction of sensitive data (credentials, tokens).
More comprehensive than application-level logging because it captures authentication events at the SDK level; more flexible than hardcoded logging because it supports multiple backends through a pluggable interface.
multi-credential-agent-support
Medium confidenceEnables a single agent to authenticate using multiple credentials (e.g., primary API key + backup certificate) and automatically fall back to alternative credentials if the primary credential fails. The SDK maintains a credential chain, attempts authentication with each credential in order, and stops at the first successful authentication. Implements a credential fallback pattern that improves resilience for agents with multiple credential sources.
Supports credential chains with automatic fallback, allowing agents to authenticate with alternative credentials if the primary credential fails. Tracks which credential succeeded and logs fallback events for audit purposes.
More resilient than single-credential authentication because it provides fallback paths; more flexible than manual credential switching because fallback is automatic and transparent to the agent.
scope-based-authorization-enforcement
Medium confidenceEnforces fine-grained authorization based on scopes embedded in the authenticated session token. The SDK extracts scopes from the Guardian's verification proof, compares requested tool scopes against authorized scopes, and rejects tool calls that exceed the agent's authorization level. Implements a scope-matching pattern that supports both exact scope matching and wildcard scope patterns (e.g., 'tools:read:*' matches 'tools:read:users' and 'tools:read:data').
Enforces authorization at the SDK level based on scopes embedded in the Guardian's verification proof, preventing unauthorized tool calls before they reach the Gateway. Supports wildcard scope patterns for flexible permission grouping.
More granular than binary allow/deny because it supports scope-based permissions; more efficient than server-side authorization checks because it enforces locally without additional round-trips.
gateway-agent-credential-exchange
Medium confidenceManages the first leg of the trilateral handshake where the Agent submits its credentials to the Gateway via an MCP tool call. The SDK encapsulates the agent's identity (API key, certificate, or token) in a standardized credential envelope, sends it through the Gateway's MCP endpoint, and receives a challenge or intermediate token. Implements credential normalization to support multiple credential types (symmetric keys, asymmetric certificates, OIDC tokens) without requiring the agent to know the Gateway's preferred format.
Normalizes heterogeneous credential types into a single envelope format before transmission, allowing agents with different credential sources to authenticate without format-specific logic. Uses MCP tool metadata to discover which credential types the Gateway accepts.
More flexible than fixed-format credential submission because it adapts to the Gateway's capabilities; simpler than building a full credential management system because it focuses only on the exchange phase.
guardian-signature-verification-delegation
Medium confidenceHandles the second leg of the trilateral handshake by forwarding the Gateway's challenge to the Guardian for cryptographic verification. The SDK constructs a verification request containing the agent's credentials, the Gateway's challenge, and metadata about the authentication context, sends it through the Guardian's MCP endpoint, and receives a signed verification proof. Implements a delegation pattern where the Guardian acts as a trusted third party that validates the agent's legitimacy without the Gateway needing to store or verify credentials directly.
Delegates verification to an independent Guardian service via MCP, creating a cryptographic proof that can be audited and replayed. The Guardian never communicates directly with the agent — all communication flows through the Gateway, maintaining network isolation.
More trustworthy than Gateway-only verification because it introduces an independent auditor; more scalable than centralized verification because Guardian can be replicated and load-balanced independently of Gateway instances.
authenticated-session-token-generation
Medium confidenceSynthesizes a final authenticated session token after both the Gateway and Guardian have validated the agent. The SDK combines the Gateway's intermediate token and the Guardian's verification proof into a single session token that encodes the agent's identity, verified scopes, and expiration time. Implements a token composition pattern where the final token is cryptographically bound to both the Gateway and Guardian's signatures, preventing token forgery or scope escalation.
Composes a single token from multiple cryptographic proofs (Gateway + Guardian), creating a non-repudiable record of which parties verified the authentication. Token includes verified scopes derived from Guardian's policies, not agent-requested scopes.
More trustworthy than single-signature tokens because it requires both Gateway and Guardian to have signed; more auditable than opaque tokens because it embeds verification metadata that can be inspected without contacting the issuer.
agent-identity-resolution-from-credentials
Medium confidenceExtracts and normalizes agent identity information from raw credentials (API keys, certificates, or tokens) into a standardized agent profile. The SDK parses credential metadata, resolves the agent's name, organization, and capabilities from a credential registry, and validates that the credential format matches the expected schema. Implements a credential-to-identity mapping that supports multiple credential sources (self-signed, CA-signed, OIDC-issued) without requiring the agent to explicitly declare its identity.
Automatically resolves agent identity from credential metadata without requiring agents to declare their identity separately. Supports multiple credential formats through a pluggable parser architecture, allowing new credential types to be added without SDK changes.
More automatic than manual identity declaration because it derives identity from cryptographic credentials; more flexible than fixed-format identity because it adapts to the credential type.
mcp-tool-schema-validation-for-auth-endpoints
Medium confidenceValidates that MCP tool schemas for authentication endpoints (credential-exchange, signature-verification) conform to the ANS specification before attempting to use them. The SDK inspects the tool's input/output schemas, parameter types, and required fields, and rejects tools that don't match the expected authentication interface. Implements a schema-driven validation pattern that prevents runtime errors from misconfigured authentication tools and provides clear error messages about schema mismatches.
Validates MCP tool schemas against the ANS specification at SDK initialization time, preventing runtime failures from misconfigured authentication tools. Uses JSON Schema validation to provide detailed error messages about which fields are missing or have incorrect types.
More proactive than runtime error handling because it catches configuration issues before authentication is attempted; more specific than generic MCP tool validation because it checks for ANS-specific requirements.
credential-type-detection-and-normalization
Medium confidenceAutomatically detects the type of credential provided (API key, certificate, OIDC token, etc.) and normalizes it into a standard format for transmission to the Gateway. The SDK uses heuristics and format analysis to identify credential types without requiring explicit type declaration, then applies type-specific normalization (e.g., PEM parsing for certificates, JWT decoding for tokens). Implements a format-agnostic credential handling pattern that reduces boilerplate for agents with different credential sources.
Uses format heuristics and cryptographic parsing to automatically detect credential types without explicit declaration, reducing boilerplate for agents with diverse credential sources. Supports multiple credential formats through a pluggable detector architecture.
More convenient than explicit type declaration because it infers type from format; more robust than regex-based detection because it uses cryptographic parsing to validate format correctness.
authentication-session-lifecycle-management
Medium confidenceManages the complete lifecycle of an authenticated session from creation through expiration and renewal. The SDK tracks session state (active, expired, revoked), enforces expiration policies, and provides hooks for session renewal before expiration. Implements a state machine pattern with explicit transitions (created → active → expiring → expired or revoked), allowing agents to proactively refresh sessions before they expire and preventing use of expired tokens.
Implements a state machine for session lifecycle with explicit transitions and renewal hooks, allowing agents to proactively refresh sessions before expiration. Provides event callbacks for session state changes, enabling agents to react to expiration without polling.
More proactive than reactive expiration handling because it warns agents before expiration; more explicit than implicit token refresh because it requires agents to opt-in to renewal behavior.
error-recovery-and-retry-logic-for-authentication
Medium confidenceImplements configurable retry logic and error recovery strategies for authentication failures at each stage of the trilateral handshake. The SDK distinguishes between transient errors (network timeouts, temporary Guardian unavailability) and permanent errors (invalid credentials, schema mismatches), and applies appropriate recovery strategies (exponential backoff for transient, fail-fast for permanent). Implements a circuit-breaker pattern to prevent cascading failures if Guardian or Gateway becomes unavailable.
Implements error classification to distinguish transient failures (network timeouts, temporary unavailability) from permanent failures (invalid credentials, schema mismatches), applying different recovery strategies for each. Uses circuit breaker pattern to prevent cascading failures.
More intelligent than blind retry because it classifies errors before deciding to retry; more resilient than no retry logic because it handles transient failures gracefully without manual intervention.
Capabilities are decomposed by AI analysis. Each maps to specific user intents and improves with match feedback.
Related Artifactssharing capabilities
Artifacts that share capabilities with @agentic-name-service/sdk, ranked by overlap. Discovered automatically through the match graph.
secure-mcp-server
MCP server: secure-mcp-server
imara
Runtime governance layer for AI agents — audit trails, policy enforcement, and compliance for MCP tool calls
@treeship/mcp
Drop-in Treeship attestation for MCP tool calls
cordon-cli
The security gateway for AI agents — firewall, auditor, and remote control for MCP tool calls
@exitmcp/core
exitMCP core: MCP server, tool registry, KV/Host/Auth interfaces
cls-mcp-server
[](https://www.npmjs.com/package/cls-mcp-server) [](https://github.com/Tencent/cls-mcp-server/blob/v1.0.2/LICENSE)
Best For
- ✓teams building multi-agent systems with distributed trust requirements
- ✓developers implementing zero-trust architectures for agent-to-service communication
- ✓organizations using MCP as their primary agent orchestration protocol
- ✓agents that make frequent MCP tool calls and need per-call authentication
- ✓teams building agent frameworks that abstract away authentication concerns
- ✓developers who want authentication to be declarative rather than imperative
- ✓organizations with compliance requirements (SOC2, HIPAA, PCI-DSS) that mandate authentication audit trails
- ✓security teams investigating authentication incidents
Known Limitations
- ⚠Requires all three parties (Agent, Gateway, Guardian) to be MCP-compatible — no fallback to REST or gRPC
- ⚠Adds latency for each authentication round-trip; not suitable for sub-100ms latency requirements
- ⚠No built-in session persistence — requires external state store for long-lived authenticated sessions
- ⚠Guardian availability is a hard dependency; no graceful degradation if Guardian is unreachable
- ⚠Only works with MCP-compatible tools; cannot route to REST APIs or gRPC services directly
- ⚠Auth context injection adds ~50-100ms overhead per tool call due to signature generation
Requirements
Input / Output
UnfragileRank
UnfragileRank is computed from adoption signals, documentation quality, ecosystem connectivity, match graph feedback, and freshness. No artifact can pay for a higher rank.
Package Details
About
Official Agent SDK for the Agentic Name Service (ANS) — orchestrates MCP tool calls across Gateway and Guardian for trilateral authentication
Categories
Alternatives to @agentic-name-service/sdk
Search the Supabase docs for up-to-date guidance and troubleshoot errors quickly. Manage organizations, projects, databases, and Edge Functions, including migrations, SQL, logs, advisors, keys, and type generation, in one flow. Create and manage development branches to iterate safely, confirm costs
Compare →Are you the builder of @agentic-name-service/sdk?
Claim this artifact to get a verified badge, access match analytics, see which intents users search for, and manage your listing.
Get the weekly brief
New tools, rising stars, and what's actually worth your time. No spam.
Data Sources
Looking for something else?
Search →